Abstract: Invasive measurement of central aortic pressure waveform(CAPW)is risky and costly. Peripheral arterial pressure waveform(PAPW)has been used to substitute CAPW. However， there is a difference between PAPW and CAPW. Although the blind system identification technology can reconstruct in real-time and dynamically estimate CAPW， it needs to be improved in optimizing the model ordering. Accordingly， a rank-based ordering method was proposed based on Hankel matrix combining the subspace method to reconstruct the central aortic pressure waveform with two peripheral pressure waveforms. The results of the validation algorithm based on the clinical data showed that the root mean square error of the measured CAPW and the estimated CAPW is about 5.87mmHg， and the waveform fit is about 74.70%. The results of the validation algorithm based on the collecting animal data showed that the root mean square error of the measured CAPW and the estimated CAPW is about 8.82mmHg， and the waveform fit is about 54.57%. Thus the experiments verified the effectiveness of the method.

Key words: aortic pulse waveform, peripheral arterial pulse waveform, blind system identification, subspace algorithm, Hankel matrix